Hopefully you mean efficiency in terms of mechanical work at the cam as a fraction of mechanical work at the crank? (Sometimes you get people who want to know what effect the pulley inertia has on engine acceleration)
You can probably get someone to provide a pretty good off-the-cuff estimate if you wait around here.
I'm guessing it'll be in the high 90s (percentage) for a simple SOHC cog/belt system.
Actually setting up a model to do a calculation would be very complex. Rule of thumb is around 90% when used at a fair percentage (say 75% or more) of rated load. If used at a very low fraction of rated load, efficiency will be much lower. If you really need an accurate number, you'll need to run experiments. Efficiency is likely to change somewhat as the belt breaks in.